sentiment-analyzer

情感分析器

分析文本内容的情感，提供详细评分、情绪检测和可视化功能。可处理单条文本、CSV 文件，或跟踪一段时间内的情感趋势。

快速开始

from scripts.sentiment_analyzer import SentimentAnalyzer

# Analyze single text
analyzer = SentimentAnalyzer()
result = analyzer.analyze("I love this product! It's amazing.")
print(f"Sentiment: {result['sentiment']} ({result['score']:.2f})")

# Batch analyze CSV
results = analyzer.analyze_csv("reviews.csv", text_column="review")
analyzer.plot_distribution("sentiment_dist.png")

功能

• 情感分类：正面、负面、中性，并提供置信度
• 极性评分：-1.0（负面）至 +1.0（正面）
• 主观性检测：区分客观与主观内容
• 情绪检测：喜悦、愤怒、悲伤、恐惧、惊讶
• 批量处理：分析包含任意文本列的 CSV 文件
• 趋势分析：跟踪一段时间内的情感变化
• 可视化：分布图、趋势图、词云

API 参考

初始化

analyzer = SentimentAnalyzer()

单条文本分析

result = analyzer.analyze("This is great!")
# Returns:
# {
#     'text': 'This is great!',
#     'sentiment': 'positive',  # positive, negative, neutral
#     'score': 0.85,            # -1.0 to 1.0
#     'confidence': 0.92,       # 0.0 to 1.0
#     'subjectivity': 0.75,     # 0.0 (objective) to 1.0 (subjective)
#     'emotions': {'joy': 0.8, 'anger': 0.0, ...}
# }

批量分析

# From list
texts = ["Great product!", "Terrible service.", "It's okay."]
results = analyzer.analyze_batch(texts)

# From CSV
results = analyzer.analyze_csv(
    "reviews.csv",
    text_column="review_text",
    output="results.csv"
)

趋势分析

# Analyze sentiment over time
results = analyzer.analyze_csv(
    "posts.csv",
    text_column="content",
    date_column="posted_at"
)
analyzer.plot_trend("sentiment_trend.png")

可视化

# Sentiment distribution
analyzer.plot_distribution("distribution.png")

# Sentiment over time
analyzer.plot_trend("trend.png")

# Word cloud by sentiment
analyzer.plot_wordcloud("positive", "positive_words.png")

CLI 用法

# Analyze single text
python sentiment_analyzer.py --text "I love this product!"

# Analyze file
python sentiment_analyzer.py --input reviews.csv --column review --output results.csv

# With visualization
python sentiment_analyzer.py --input reviews.csv --column text --plot distribution.png

# Trend analysis
python sentiment_analyzer.py --input posts.csv --column content --date posted_at --trend trend.png

CLI 参数

参数	描述	默认值
--text	要分析的单条文本	-
--input	输入 CSV 文件	-
--column	文本列名称	text
--date	用于趋势分析的日期列	-
--output	输出 CSV 文件	-
--plot	保存分布图	-
--trend	保存趋势图	-
--format	输出格式（json、csv）	json

示例

产品评价分析

analyzer = SentimentAnalyzer()
results = analyzer.analyze_csv("amazon_reviews.csv", text_column="review")

# Summary statistics
positive = sum(1 for r in results if r['sentiment'] == 'positive')
negative = sum(1 for r in results if r['sentiment'] == 'negative')
print(f"Positive: {positive}, Negative: {negative}")

# Average sentiment score
avg_score = sum(r['score'] for r in results) / len(results)
print(f"Average sentiment: {avg_score:.2f}")

社交媒体监控

analyzer = SentimentAnalyzer()

# Analyze tweets with timestamps
results = analyzer.analyze_csv(
    "tweets.csv",
    text_column="tweet_text",
    date_column="created_at"
)

# Plot sentiment trend
analyzer.plot_trend("twitter_sentiment.png", title="Brand Sentiment Over Time")

客户反馈分类

analyzer = SentimentAnalyzer()

feedback = [
    "Your support team was incredibly helpful!",
    "The product broke after one day.",
    "Shipping was on time.",
    "I'm extremely disappointed with the quality.",
    "It works as expected, nothing special."
]

for text in feedback:
    result = analyzer.analyze(text)
    print(f"{result['sentiment'].upper():8} ({result['score']:+.2f}): {text[:50]}")

输出格式

JSON 输出

{
  "text": "I love this product!",
  "sentiment": "positive",
  "score": 0.85,
  "confidence": 0.92,
  "subjectivity": 0.75,
  "emotions": {
    "joy": 0.82,
    "anger": 0.02,
    "sadness": 0.01,
    "fear": 0.03,
    "surprise": 0.12
  }
}

CSV 输出

text	sentiment	score	confidence	subjectivity
优质产品！	positive	0.85	0.91	0.80
糟透了……	negative	-0.72	0.88	0.65

依赖项

textblob>=0.17.0
pandas>=2.0.0
matplotlib>=3.7.0

局限性

• 针对英语进行了优化（其他语言的准确度可能较低）
• 可能无法准确检测讽刺和反语
• 可能会遗漏依赖上下文的情感
• 短文本（少于 5 个词）的置信度较低